PROJECT ABSTRACT Cataracts result from progressive aggregation of eye lens crystallin proteins. Severity of age-onset cataract has been linked to specific post-translational modifications in crystallins that accumulate during aging. Two important classes of cataract-associated modifications are oxidation of Trp residues to more hydrophilic products and oxidation of Cys residues to generate disulfide bonds. Our prior research has revealed a crucial synergy between the two. The W42Q variant of human ?D crystallin (a Trp oxidation mimic) and the W42R congenital-cataract variant, are destabilized but well folded and soluble under reducing conditions, yet formation of a non-native internal disulfide bond (Cys32-Cys41) kinetically traps them in a partially unfolded conformational intermediate, generating rapid and robust aggregation at physiologically relevant temperature, pH, and concentration in vitro. We have developed a rapid, atomistic Monte-Carlo modeling method, with a knowledge-based statistical potential, that is uniquely suited for the study of conformational intermediates, including in multiple polypeptide chains as they simultaneously unfold to reveal new protein-protein interactions. We have already applied this method to ?D crystallin and its variants to predict not only the structure of the aggregation-prone intermediate but also, for the first time, an atomistic model of the aggregated state. Experimentally, we recently discovered a novel oxidoreductase activity in human ?D crystallin and demonstrated that native-state disulfides in WT can be transferred to generate the non-native, aggregation-promoting disulfide in W42Q. We found an even more surprising WT/mutant interaction ? domain interface stealing ? that allows WT to catalyze mutants? aggregation even in the presence of an abundant external disulfide source. We will now (1) investigate the physical principles, kinetics, and evolutionary and disease implications of the novel interface stealing interaction by a combined computational, biochemical, and proteolysis/mass spectrometry approach we are now developing; and (2) distinguish among atomistic models for the aggregation precursor and the aggregated state and (3) apply these newly refined atomistic models to rationally design structure-based peptide inhibitors of the aggregation process. Although our studies have focused on the W42Q/R variants, other cataract-associated variants (V75D, L5S) appear to behave quite similarly. Moreover, both the native and the non-native disulfide we identified as culprits in aggregation processes have been entirely supported by tissue proteomics of aged and cataractous human lenses in the absence of any genetic mutation. We will therefore test the hypothesis that many mutations or post- translational modifications converge on few conformational intermediates that determine aggregation. We will generalize the detailed mechanistic and structural picture of aggregation to other ?-crystallins and other cataract- associated variants testing whether human ?C and ?S crystallins are also redox-active and capable of interface stealing. A more general understanding of the synergy between structural destabilization and redox chemistry in cataract will improve design of aggregation inhibitors testable on existing genetic mouse models of cataract.